Monitoring of cytosol redox potential in wild-type plants expressing c-roGFP1 (Fig. 4A) revealed that on day 14 of water deprivation the redox potential increases, becoming more oxidized in both young and mature WS leaves, reaching its most oxidized level on the maximum stress day (day 18), when WS plants presented values of RWC ca. 60%, a 15% lower than WW plants. The mature (M) and young (Y) leaves clearly have different ranges of redox potential: in the cytosol of WW plants values where found between −307 and −303 mV in M leaves and below −310 mV (less oxidized) in Y leaves. Under maximum WS, wild-type M leaf cytosol reached a maximum of −295 mV, while Y leaves reached only −299 mV. With re-watering and recovery of WS plants, the redox potential at day 24 decreased back to −306 and −312 mV (a more reduced status) in M and Y leaves, respectively. Subsequently, mild water stress induces a reversible oxidative stress. Moreover, the redox potentials of WW plants show a tendency to become less negative (more oxidized), through the course of the experiment and showed a relationship between leaf age and redox potential.

For vtc2 plants values for the cytosol redox status (Fig. 4B) in WW plants showed similar trends as observed in the wild-type, although, unexpectedly under WS, redox values appear lower (more negative) and thus less oxidized than in the wild-type cytosol and also differences between leaf typology were less noticeable. Mutant vtc2 plants reached a maximum potential cytosolic redox value of −301 mV in M leaves cytosol on day 18 of WS and a maximum of −307 mV in the cytosol of Y leaves. With re-watering, recovery of the redox potential of the WS plants was similar to that of WW wild-type plants. This suggests that a ascorbate levels in the vtc2 plants, though 80% less than wild-type, may still be adequate to buffer WS-induced redox changes. Alternatively, another mechanism, not related with the levels of ascorbate, may function to protect vtc2 mutants from intensive changes in oxidative redox status.

Mitochondria redox potential was also monitored in wild-type and vtc2 plants (Fig. 5). For WW Col-0 plants the mitochondrial redox status was more negative (more reduced) than in the cytosol (Fig. 5A) for both types of leaves. Under WW conditions the mitochondrial redox potential of wild-type mature leaves became more oxidized, from −360 mV to −351 mV with plant age, (Fig. 5A). However, the effects of age along the experiment are not observed in young leaves, which remain more reduced compared with mature leaves, showing values of ca. −358 mV.

For the vtc2 mutant (Fig. 5B) similar values and trends were also observed; in WW mature leaves mitochondrial redox potential increases (becomes more oxidized) with plant age, from a most reduced value of −359 mV on day 4, to a more oxidized values of −352 mV at the end of the experiment. Under WS treatment the redox potentials for mitochondria in vtc2 plants showed remarkable differences between M and Y leaves; while redox potential increases (becomes more positive) with WS in M leaves (reaching a maximum of −349 mV on day 18), in Y leaves remains constant during the entire time course of stress (ca. −360 mV). Since these trends are the same for both Col-0 plants and vtc2, differences in redox potentials are therefore related to leaf age. Six days after beginning watering recovery the redox potential of WS plants recovered to more reduced values of −354 and −359 mV for old and young leaves, respectively.

Total leaf ascorbate content (AsA+DHA) was determined during time course of water deprivation and recovery in M and Y leaves of wild-type and vtc2 transformed lines. In order to ascertain whether expression of roGFP1 affected total ascorbate content, ascorbate levels were measured and compared in roGFP1-transformed and non-transformed wild-type and vtc2 plants under both WW and WS conditions. No differences were noted in ascorbate content so we conclude that being transformed with roGFP1 does not affect ascorbate levels.

In WW Col-0 plants total ascorbate leaf content ranged from 1 to 1.9 μmol g⁻¹ FW in M leaves and from 1.7–2.8 μmol g⁻¹ FW in Y leaves (Fig. 6A). Under WS treatment total ascorbate increases in comparison with the WW treatment, but with M leaves showing significant lower ascorbate content than Y leaves (Fig. 6A and B). DHA increases successively, both on day 14 and 17 showing a significative increase prior to water recovery and significantly decreases during WS recovery (Fig. 6B).

Total ascorbate levels in vtc2 mutants were about 80% lower than in Col-0 plants (Fig. 6C). Young leaves had significantly higher levels of ascorbate than mature leaves (about 50% more) and higher DHA levels during WS that significantly decrease after re-watering (Fig. 6D).

Leaf glutathione was also determined from the same extract used for measurements of ascorbate (Fig. 7). No significant differences were observed in the Col-0 ecotype between WW and WS treatments. However, glutathione content in Y leaves (ranging from 0. 37–0. 57 μmol g⁻¹ FW) was higher (a 2–8% increase) compared with M leaves (ranging from 0.26–0.37 μmol g⁻¹ FW) during the time course in both treatments (Fig. 7A and B). During recovery the fast answer of Y leaves was shown by a significant increase after 6 h (19 d) and back to previous values after 6 d (19 d) (Fig. 7B).

The levels of total glutathione in vtc2 mutants tend to increase through the experiment in WW plants (about 40% at day 19 above day 0) and also total glutathione content increases (4–10%) in young compared with mature leaves (Fig. 7C and D). Through the duration of the experiment, and up until after 6h re-watering, glutathione increases in both Y and M leaves of water-stressed plants, reaching values about 40% higher than day 0. One week after re-watering and recovery (day 24) the levels of total glutathione decrease to values of 0.35 μmol g⁻¹ FW in both M and Y leaves, similar to values observed at the beginning of the experiment. The peak of glutathione after 6h of re-watering was not only significant in Y leaves but also in M (by almost 2-fold higher than maximum stress values).

Water stress results in a reduction in biomass that becomes significantly different from WW plants in day 10 for the Col-0 but only after water recovery for vtc2 line. During the water withholding, the total biomass for the vtc2 line was identical for both the WW and WS plants. This suggests that a reduction in ascorbate levels in the vtc2 mutant did not impair the ability of this line to respond to WS, at least up to day 18, following the same growth ratio as WW plants. Moreover, at day 24 the biomass for WS vtc2 line is only 16% lower than for WS Col-0 and, again supporting the view that the reduced levels of ascorbate in vtc2 possibly are not linked to the reduction in biomass due to WS. However, the fact that the biomass of the WW vtc2 plants is significantly less at day 24 compared with the biomass of the WW Col-0 plants, suggests that a reduced level of ascorbate in WW vtc2 plants is not without consequence and thereby points to a role for ascorbate in plant growth under non-stressed conditions.³⁴

Water stress affects the redox status differently in the Col-0 vs. the vtc2 line. In the Col-0 line the most oxidized status of the mitochondria is reached 14 d after withholding water for both Y and M leaves, whereas in the vtc2 mutant the mitochondria compartment was at its most oxidized just before re-watering (18 d after beginning to withhold water). On the other hand, cytosol redox values for the Col-0 line were the most oxidized just before re-watering and attained values of redox potential less negative (more oxidized) than mitochondria. Our results show that mitochondria on both Col-0 and vtc2 lines are better buffered than are the cytosol. These results are in accordance with Jiang et al.¹² who showed that mitochondria are much better than cytosol in buffering changes in redox potential

The redox status of the mitochondria compartment also appears related to leaf age, with younger leaves of both the Col-0 and vtc2 lines consistently exhibiting a more reduced redox status than older, more mature leaves. But with regard to the redox status of the cytosol compartment, its relationship between leaf age is less pronounced. From this work we conclude that the mitochondrial and cytosolic compartments respond differently to WS, with regard to redox status.

Moreover, because no significant differences were seen in mitochondria redox potentials in young leaves between WW or WS, this again points to the high buffering capacity of mitochondria, especially in young leaves.

It is known that mitochondria are highly resistant to abiotic stress³ which our work now suggests could be related to their high capacity of redox homeostasis. Additionally, because mitochondria can remain intact until the last stages of senescence¹⁵ and this could explain our observation in WS mature leaves mitochondria remain more reduced than the cytosol.

In young leaves of WW plants (Col-0 and vtc2), there are no significant differences in redox status for the duration of the experiment at the mitochondria. Because mature leaves of WW plants tend to be more oxidized throughout the experiment, this increase in ROS could underlie the leaf senescence process.²⁰ Next, we consider changes in redox status in relation to variations in the levels of ascorbate and glutathione.

Both the oxidized and reduced forms, and total amounts, of ascorbate and glutathione change in response to WS. Most pronounced are changes in AsA and DHA in young leaves, and to a lesser extent, in mature leaves, after 14 d withholding water, with (total ascorbate peaking at day 18). Even after re-watering, AsA levels continue to rise (on day 19), though total measured ascorbate (oxidized and reduced forms) is somewhat less than on day 18. To some degree this same pattern is observed in the WS vtc2 line with a peak at day 17, when, however, in young leaves reduced ascorbate is undetectable, but still with relatively high levels of DHA. This suggests that the prolonged WS efficiently causes oxidation of the reduced form (AsA) to DHA and/or negatively affects the regeneration of AsA. For WS vtc2 plants, the ascorbate buffering could still have been just enough to buffer WS-induced redox changes. Alternatively, another mechanism, not related with the levels of ascorbate, may function to protect vtc2 mutants from extensive changes in oxidative redox status. With re-watering the level of AsA increases (DHA levels remain unchanged compared with day 17), suggesting that WS may also affect AsA synthesis. In summary, while there are many subtle differences in AsA/DHA levels induced by WS, and linked to leaf age, the most pronounced effects on ascorbate levels under WS were observed in young leaves. In parallel, young leaves showed a more reduced redox potential than mature leaves. Moreover, the last steps of ascorbate biosynthesis are located in mitochondrial inner membrane^2,18,30 and thus the increase in ascorbate biosynthesis in Col-0 Y leaves would explain the low redox potential in mitochondria throughout WS treatment.

Unlike the situation for ascorbate, WS results in less significant, if any, changes in glutathione (GSH and GSSG). In both the Col-0 and the vtc2 lines GSH and GSSG amounts remain relatively unchanged through day 18 and only increase after re-watering. This suggests that these plants do not rely on changes in GSH/GSSG to buffer WS, at least through day 18, and also that because re-watering results in an increase in total glutathione (notably in the young vtc2 leaves), that glutathione synthesis may be impaired by prolonged water stress. Additionally, even in the ascorbate-deficient mutant, glutathione levels remain more or less unchanged in WS plants, pointing to the fact that no increase glutathione occurs to compensate for a reduction in ascorbate. In addition, though mitochondria appear to be heavily buffered by glutathione,^17,35 the increasingly oxidized redox status, with the prolongation of WS, can be explained as a consequence of no change in GSH/GSSG levels. On the other hand, because changes in the redox status of the mitochondria (which start to become more reduced after day 14) parallel increases in ascorbate, this suggests that in the mitochondria ascorbate is the major redox buffering species. This view is supported by essentially no change in mitochondrial redox status in the ascorbate-deficient mutant.

The results observed in AsA levels in the Arabidopsis Col-0 line plants subjected to water stress and cultivated in a growth chamber cannot now be extrapolated in its totality to plants subjected to drought and growing under field conditions. Nevertheless, our results suggest that even in plants growing in natural conditions and subjected to a progressive water stress, that variations in the endogenous concentrations of AsA will likely be useful as stress markers, since plants resistant to drought generally show an increase in AsA levels.¹⁴

In summary, the results of measurements of ascorbate and glutathione accord well with the roGFP data. Most notable is the fact that the redox status of cytosol of WS Col-0 plants continues to become more oxidized with prolongation of WS, whereas the redox status of the mitochondria starts to become more reduced at about day 14. We therefore conclude that Arabidopsis responds to WS by increasing the synthesis of ascorbate, which primarily buffers changes in the mitochondria. Second, WS appears not to result in a significant increase in glutathione synthesis (or in the amounts of GSH or GSSG). Thus, if ascorbate is the major redox-buffering species in water-stressed plants, and if its reduced form normally involves the glutathione cycle, then the increase in AsA/DHA in WS plants may occur via other reductants, in addition to GSH.

Seeds of Arabidopsis thaliana Columbia (Col) ecotype wild-type (Col-0) and the ascorbate deficient mutant (vtc2) were transformed with a redox-sensing green fluorescent protein [reduction-oxidation-sensitive green fluorescent protein] targeted either to the cytosol (c-roGFP1) or to mitochondria (m-roGFP1) using the floral dip method.^4,12 Selected seeds expressing c-roGFP1 and m-roGFP1 were surface sterilized and plated on MS agar medium under sterile conditions and then incubated for 4 d at 4°C whereupon they were transferred to a growth chamber.

At the stage of four true leaves the seedlings were transplanted to a substrate mixture of peat/perlite/vermiculite (3:1:1; v/v/v) and maintained in a growth chamber (8 h. photoperiod; PPFD, 110 μmol m⁻² s⁻¹, 23.5°C). After 4 wk the experiment commenced (day 0) by imposing two watering regimes: (1) Plants were watered to saturation with Hoagland solution twice a week [substrate at approximately 90% of relative water content (RWC)] throughout the experiment [well-watered (WW) plants] and (2) water was withheld from plants [water-stressed (WS) plants] during 18 d, performing on day 18 the re-watering of these plants.

After 10 d of sampling (5-wk-old plants) two types of leaf morphologies were distinguished on plants: (M) mature; which are the first produced leaves, and which were smaller and lacking trichomes; and (Y) young developing leaves which are larger and showed a high number of trichomes per unit area compared with mature leaves.³² These typologies were taken in account when making ratiometric measurements as well as in the antioxidant analysis.

After 18 d of withholding water the WS plants were re-watered to saturation and allowed to recover under the same conditions under which the WW were maintained. A diagram of the experimental design is shown in Figure 8.

Ratiometric measurements were made on 5-wk-mature WS and WW plants as described by,¹³ keeping track of which leaves were M or Y. Without detaching from the plant, two representative leaves of both established types were selected from each plant and directly mounted dry on a microscope slide and covered dry with a cover glass. Measurements were made according to Jiang et al.¹²

Briefly, a Nikon Diaphot Nikon FN600 microscope (Nikon) fitted with Plan Fluor 10 × /0•30 N.A. dry objectives and a Chroma Technology Phluorin filter set (exciters D410/30 and D470/20, dichroic mirror 500DCXR, emitter HQ535/50) microscope was used. Image acquisition and processing were performed using a Hamamatsu Orca-100 cooled CCD camera (Hamamatsu Corp.) and emission data (fluorescence) were collected over time using MetaFluor 6.1 image analysis software (Molecular Devices) which controlled both filters and the data collection. Fluorescence ratios were obtained by dividing the intensities obtained at 410 nm and 474 nm (i.e., 410/470). Each image was corrected for background by subtracting the intensity of an adjacent cell-free region. The 410/470 fluorescence ratios were normalized by using maximal reduced and oxidized values obtained after perfusing the leaf sequentially with adding 1 M H₂O₂ and 500 mM DTT, in order to not only drive the ro-GFP toward the fully oxidized and reduced forms, respectively, but also to prove the reversibility of the indicator. Images of these selected leaves were taken and saved within 5 min or less and then H₂O₂ and DTT were added to attain the normalized results.

The maximal ratio under maximally oxidized conditions was set equal to 1.0 and the minimum ratio measured during maximally reduced conditions was set equal to 0.0. These normalized fluorescence ratios were then converted to redox potentials using the calibration curve generated.¹¹

Their experimentally determined data for fluorescence ratio values obtained from isolated recombinant roGFP1 titrated with DTT_red to DTT_ox solutions with calculated redox potentials were fit using a nonlinear regression equation:

and the resulting curve (y₀ = 0.01, a = 0.99, x₀ = −288 and b = 13) was then used to calibrate these experiments.

For each measurement at least six different plants were used. From each of these plants one leaf was chosen for the redox potential measurements and the remaining leaves were used for the water relations. For processing the image data, four areas in the leaf, were selected to measure the fluorescence ratio.

The aerial part of the rosettes was used for both redox measurements and to assess plant water status. RWC (%) = (FW−DW)/(TW−DW), where FW is the fresh weight, TW is the turgid weight after rehydrating the samples for 24 h at 4°C in darkness and DW is the dry weight after oven-drying samples at 80°C until constant weight.

Gravimetric soil water content (SWC) was directly determined using the difference in soil weight immediately after the sampling (FW) and after oven drying (DW).

For ascorbate and glutathione determinations, 100 mg of leaves were ground in liquid nitrogen with a mortar and pestle, followed by addition of 1ml of extraction buffer (6% meta-Phosphoric acid; 2.5 mM EDTA) and clarified with 10 min centrifugation at 8,500 g. Re-extraction using the same conditions was performed and the two supernatants were mixed. Finally, extracts were neutralized and conveniently diluted before spectrophotometric readings, following.²¹ The reduced and oxidized forms of both ascorbate and glutathione were determined by using a 96 well quartz microplate (Hellma) and plate-reader, as described in.²⁸ The levels of ascorbate (AsA) (reduced) and dehydroascorbate (DHA) (oxidized) were determined using ascorbate oxidase (AO) and dithiothreitol (DTT), respectively, as described in.⁶ Ascorbate oxidase specifically oxidizes all reduced ascorbate in the sample. Thus the decrease in O.D. at 265 nm is related to AsA concentration. Alternatively, when the samples are incubated with DTT, DHA is reduced to AsA and the increase in O.D. is proportional to the initial DHA concentration.

Total glutathione was determined using the glutathione reductase (GR) recycling assay,³³ while for oxidized glutathione determination samples were incubated with 2-vinylpiridine.¹⁰

The data were analyzed using PASW for Windows v. 17.0.2 (SPSS Inc.). ANOVA was used to compare mean values at a range of sampling times for RWC, glutathione and ascorbate oxidative status and content and ratiometric measurements of redox potential. The post-hoc Duncan’s test was applied. Significance levels of 95% (p < 0.05) are indicated in figure legends. At each sampling time significantly different means are marked with different letters.